The Reciprocal Effects of Brand Equity and Trivial Attributes

Journal of Marketing ResearchVol. XL (May 2003), 161–175161

*Susan M. Broniarczyk is Associate Professor of Marketing, McCombsSchool of Business, The University of Texas at Austin (e-mail: [email protected]). Andrew D. Gershoff is Assistant Professorof Marketing, Graduate School of Business, Columbia University (e-mail:[email protected]). The authors thank Kevin Keller, Joe Urbany, andthe JMR reviewers for their valuable comments. The authors are listedalphabetically.

SUSAN M. BRONIARCZYK and ANDREW D. GERSHOFF*

Brands increasingly introduce products with attributes that fail to pro-vide consumers with meaningful benefits (i.e., trivial attributes). Theauthors present two experiments that examine the effect of brand equityon consumer valuation of such trivial attributes and the reciprocal effectthat such a strategy may have on brand equity. The results show that bothhigh and low equity brands benefit from offering an attractive trivial attrib-ute in the absence of a disclosure of its true value. However, prechoicedisclosure of an attribute’s triviality heightens the role of the brand andcontext cues. Competing low equity brands benefit by sharing the trivialattribute with a higher equity brand, whereas competing high equitybrands benefit from uniquely offering a trivial attribute. Postchoice reve-lation that an attribute is trivial hurts the subsequent ability of a low butnot a high equity brand to differentiate in a new product category, partic-ularly among subjects who had previously chosen the target brand. Forinsights on brand dilution, the authors also examine consumer

attributions regarding marketer intent for offering a trivial attribute.

The Reciprocal Effects of Brand Equity andTrivial Attributes

Trivial differentiation occurs when a brand differentiateson an attribute that “appears valuable but, on closer exami-nation, is irrelevant to creating the implied benefit” (Car-penter, Glazer, and Nakamoto 1994, p. 339). For example,Pantene Pro-V shampoo differentiates from other shampooson the basis of its provitamin ingredients. Consumers maybelieve that because the ingestion of vitamins improvesoverall health, vitamins have a similar positive effect whenapplied externally to hair. Yet according to ConsumerReports (2000, p. 19), vitamins in shampoo have no benefi-cial effect on hair (which is dead cells) and “really onlyenliven one thing: the label copy.” Similarly, the Epson GT-1200 scanner offers interpolated resolution that estimatesdata points between those that are measured, seemingly toenhance image equity. But “impressive-sounding specifica-tions like ‘interpolated resolution’ … [are] hype, overkill,plain and simple. In truth, there’s only one kind of resolutionthat really matters—optical resolution, which is derivedfrom the number of sensing elements” (Consumer Reports2001, p. 26).

Evidence suggests that even when consumers are pre-informed of a trivial attribute’s lack of performance value,the attribute continues to have a positive impact on con-sumer choice (Carpenter, Glazer, and Nakamoto 1994). Wepropose that the ability of a brand manager to capitalize onthe use of a trivial attribute strategy to create a competitiveadvantage is dependent on a brand’s equity. Specifically, ourresults show that when the value of a trivial attribute iscalled into question, low equity brands benefit over closecompetitors if they share the attribute with a higher equityalternative, whereas high equity brands benefit if theyuniquely offer the attribute.

Although brand equity can be leveraged through a trivialattribute strategy, doing so may carry a sizable downside. Theuse of trivial attributes risks undermining the credibility of thebrand, which serves as the foundation of its equity (Erdemand Swait 1998). Detrimental effects may be exacerbated ifconsumers purchase a brand in good faith only to discoverthat the differential was trivial. Ramifications may includeconsumers reacting negatively to the brand’s other productsand its future differentiation. For example, if after purchasingan Epson scanner, consumers learn from Consumer Reportsthat interpolated resolution provides no advantage, Epsonmay be penalized when it offers other new scanner attributes,as well as when it offers new attributes in other categoriessuch as printers. Our results show that such brand dilution ismore likely for low than for high equity brands.

In this article, we first review the trivial attribute literatureand then hypothesize about the role of brand equity on con-

162 JOURNAL OF MARKETING RESEARCH, MAY 2003

sumer valuation of trivial attributes, as well as about theconsequences of a trivial attribute strategy in future evalua-tions of a brand. We present two experiments that examinethe relative competitive advantage that a low versus highequity brand can gain in its equity tier (i.e., intra–equity tiercompetition) from using a trivial attribute strategy. Studies 1and 1A examine the effect of prechoice disclosure of theattribute’s true value on consumer valuation of the trivialattribute. We show that positive valuation of the trivialattribute is dependent on brand equity and the uniqueness ofthe attribute relative to the context set. Study 2 varies thetiming of the disclosure in an initial category as either pre-or postchoice and then examines its effect on subsequentconsumer reaction to a brand’s new differentiated offering ina related product category. When disclosure of the trivialattribute occurs after an initial product choice, valuation ofa brand’s subsequent differentiation depends on the sub-jects’ initial choice and the brand tier.

TRIVIAL ATTRIBUTES

Brown and Carpenter (2000, p. 372) define trivial attrib-utes as “those attributes with a trivial and/or subjective rela-tionship to perceived quality as well as objectively irrelevantattributes.” Trivial attributes include those that provide noperformance benefit but that consumers may perceive as“ambiguously positive” (e.g., “alpine class” versus “regular”down fill) (Brown and Carpenter 2000, p. 374) or for whichconsumers may have existing preferences (e.g., silk as aningredient in shampoo) (Carpenter, Glazer, and Nakamoto1994). In addition, trivial attributes include fictional attrib-utes that may provide novel associations without adding anyobjective benefit to the product (e.g., “Fahrvergnuegen” todescribe Volkswagen) (Brown and Carpenter 2000; Mukher-jee and Hoyer 2001). Trivial attributes may also includethose attributes that have known value to consumers but areassumed to be irrelevant to both the brand performance andthe consumer choosing the product (e.g., undesired promo-tional premiums) (Brown and Carpenter 2000; Simonson,Carmon, and O’Curry 1994; Simonson, Nowlis, and Simon-son 1993). We limit our study of trivial attributes to thosesituations in which a brand offers an attribute level that isdistinct from its competitor, for which consumers may havea prior preference, but provides no meaningful performancebenefit.

Although consumers may value attractive trivial attrib-utes, in practice the long-term viability of a trivial attributestrategy seems less than tenable. Market forces such as com-petitive advertising or consumer advocate groups might pro-vide consumers with explicit objective information aboutthe attribute’s irrelevance, reducing the ambiguity of itsvalue. However, research suggests that such disclosure maynot attenuate the trivial attribute’s advantage. In a strong testof trivial differentiation, Carpenter, Glazer, and Nakamoto(1994) show that disclosure of the irrelevance of a trivialattribute before product exposure does not by itself elimi-nate the effect of its differentiating value in consumers’product judgments. Subsequent research has shown that theelements of the decision context affect valuation of a trivialattribute.

In decision contexts in which there is no other diagnosticinformation, consumers have been shown to temporarilytreat trivial attributes as though they have value to strategi-

1Carpenter, Glazer, and Nakamoto’s (1994) study consists of three pricepoints described as low, high, and premium. We use a nine-point scale toevaluate brands, and our three levels of brand equity were described as low(x� = 3.5), mid (x� = 5.0), and high (x� = 7.5). Thus, we match our high–brandequity result to Carpenter, Glazer, and Nakamoto’s (1994) premium-pricelevel result. It may have been possible to find brands rated higher than ourhigh tier.

cally resolve the dilemma of choosing between otherwisecomparable alternatives. Brown and Carpenter (2000) showthat valuation of the trivial attribute is dependent on its abil-ity to expedite a final choice and not necessarily on anunderlying belief that it improves product performance.Specifically, in choice sets of more than two comparablealternatives, positive valuation of a trivial attribute is morelikely because it allows for a simple resolution to the choiceproblem, whereas a negative valuation would leave a diffi-cult choice from the remaining alternatives.

In decision contexts in which other diagnostic productinformation is available, motivated consumers are likely touse it to infer a value to the trivial attribute (Feldman andLynch 1988). For example, consumers have been shown torely on the labels of trivial attributes to make inferencesabout the value of the attribute (Broniarczyk and Gershoff1997). Of particular interest, in Carpenter, Glazer, andNakamoto’s (1994) second study, consumers were shown touse price information to form inferences about the trivialattribute. In the absence of a disclosure regarding the trivialattribute’s null value, a positive linear relationship existedby which consumers inferred that higher (lower) prices wereassociated with a higher (lower) value for the trivial attrib-ute. However, on disclosure of the trivial attribute’s irrele-vance, the consumer decision-making process became morecomplex. At the moderate price level, the positive benefitfrom trivial differentiation continued, whereas no effect wasobserved at the low price level. However, the disclosureattenuated consumer valuation of the trivial attribute at apremium price.

Because price level and brand equity are often positivelyrelated (Chaudhuri and Holbrook 2001), Carpenter, Glazer,and Nakamoto’s (1994) results might imply that trivial dif-ferentiation is an ineffective strategy for high equity brandsin the presence of a disclosure and for low equity brands ingeneral.1 However, key differences between brand and pricecues suggest that a trivial attribute strategy may be effectivefor both high and low equity brands. First, brand cues are anenriched set of associations, whereas price cues are refer-ence dependent (Nowlis and Simonson 1997). Second, con-sumers may infer that companies are motivated to maintaintheir brand reputations in the long run for future transactions(Shapiro 1982). Third, although a price cue may signalincreased value, it also has negative consequences associ-ated with higher costs to the consumer. A premium-pricedalternative that achieves increasing profits without providinga similar increase in value may be less preferred by con-sumers if they ascribe a negative motive to the alternative(Campbell 1999; Kahneman, Knetsch, and Thaler 1986).Campbell (1999) provides evidence that consumers reactdifferently to brand cues by showing that high brand equitycan attenuate the likelihood of such negative inferences fol-lowing a price increase.

Brand Equity and Trivial Attributes 163

2Nowlis and Simonson (1996) show that the effect of a new product fea-ture on brand choice is a function of multiattribute diminishing sensitivityand performance uncertainty. A high quality, high priced brand benefitsmore than a low quality, low priced brand from the addition of a new fea-ture. Given diminishing returns of additional differentiation, it is unlikelythat a high quality brand would risk the potential negative repercussions ofa trivial attribute strategy to further distance itself from a low quality brand.

3Discovering the truth about the value of the differentiation requires con-siderable consumer motivation and the opportunity and ability to integratethis information with existing knowledge (Johar and Simmons 2000).

BRAND EQUITY AND DISCLOSURE OF TRIVIALATTRIBUTES

Keller (1993) defines brand equity as the differentialeffect of brand knowledge on consumer response to thebrand’s marketing activities. Brand knowledge is the set ofassociations that consumers hold in memory regarding thebrand’s features, benefits, users, perceived equity, and over-all attitude as a result of prior brand marketing activities andinvestments in advertising and promotion. Brand knowledgehas been shown to differentially affect consumers’ responsesto blind taste tests (Allison and Uhl 1964), subjective per-ceptions of attribute values (Park and Srinivasan 1994), andinferences about the attributes of brand extensions (Park,Milberg, and Lawson 1991).

Our primary interest is how consumers use brand knowl-edge as a cue to infer the value of a trivial attribute and howthe use of trivial attribute strategies affects subsequent brandequity. We consider situations in which brands employ triv-ial attribute strategies to gain advantage over their closecompetitors that have comparable levels of equity and oth-erwise offer similar attributes. We limit our focus to suchintra–equity tier competition for two reasons. First, the pres-ence of meaningful differentiation has been shown to domi-nate a trivial attribute in product evaluation when consumersare informed of its true value (Broniarczyk and Gershoff1997). We expect that a trivial attribute strategy will notdominate the relevant attribute of brand equity and thereforewill not result in a low equity brand being preferable to ahigh equity brand.2 Second, with such intratier competition,we are able to isolate how brand equity affects the valuationof the trivial attribute.

Although we examine the subjective evaluation of trivialattributes, we focus on the situation in which the trivialattribute’s null value is explicitly disclosed. Marketers willstrategically label a trivial attribute so that it gives theappearance of being valuable, which leads to positive sub-jective consumer inferences. These inferences from theattribute label are likely to be generated spontaneously,according to subjects’ prior beliefs (Broniarczyk and Alba1994), and only if subjects engage in further processing willthe validity of the attribute be questioned (Gilbert 1991).3Thus, in the absence of a disclosure of its true value, weexpect that both low and high equity brands benefit over anintratier competitor by offering an attractive trivial attribute.

However, when the triviality of the attribute is disclosed,consumers must reconcile why the brands in the choice setdifferentiate the attribute at all (Carpenter, Glazer, andNakamoto 1994). Given this uncertainty, we expect con-sumers to go beyond the trivial attribute’s label and engagein more constructive information processing (Houston,Childers, and Heckler 1987). Such constructive processingshould make them more sensitive to brand equity and con-

text cues (Buchanan, Simmons, and Bickart 1999) whenthey infer value for the trivial attribute.

Brand Equity

Brand equity provides information on the value of thetrivial attribute in addition to the attribute label and disclo-sure. In the case of high equity brands, prechoice disclosureis countered by consumers’ strong favorable performanceassociations for the brand, which leads to positive inferencesregarding the trivial attribute. In the case of low equitybrands, prechoice disclosure confirms consumers’ lowexpectations regarding the performance of the brand’sattributes.

Furthermore, brand equity serves not only to create spe-cific associations in consumers’ minds but also as a generalmarket signal of the credibility of these brand associations(Erdem and Swait 1998). This argument is supported byprior research that shows that positive brand equity providesgoodwill value in the face of uncertainty (Shapiro 1982).The credibility of high relative to low equity brands has ledconsumers to be more likely to believe their extreme adver-tising claims (Goldberg and Hartwick 1990) and longer war-ranty claims (Boulding and Kirmani 1993). Compared withlow equity brands, high equity brands reduce the negativeinfluence on consumer choice of an unattractive sales pro-motion (Simonson, Carmon, and O’Curry 1994) and reduceconsumers’ negative inferences about marketer motives aftera price increase (Campbell 1999).

The preceding suggests that high equity brands are morelikely than low equity brands to benefit from offering a triv-ial attribute when its irrelevance is disclosed. Pertinent to therole of disclosures, advertising to correct deceptive adclaims (i.e., corrective advertising) has been shown to atten-uate the negative effects on product evaluation for high rel-ative to low reputation brands (Johar 1996). A high brandequity cue might be expected to have a particularly strongmitigating effect on a report of irrelevance of a trivial attrib-ute because the disclosure does not imply a negative per-formance value for the trivial attribute (Ditto and Lopez1992). Rather, it informs consumers of the trivial attribute’snull effect and suggests that they should be indifferent to itduring their product deliberations. However, the uncertaintyraised by the disclosure may lead consumers to consult otherinformation in the decision context.

Decision Context

Contextual information provided by intertier competitionis also a source for inferences regarding a trivial attribute(Prelec, Wernerfelt, and Zettelmeyer 1997; Simonson,Nowlis, and Lemon 1993). Carpenter, Glazer, andNakamoto (1994) show the benefit garnered by a singlealternative that uniquely offers a trivial attribute in a set ofcontext alternatives. However, multiple brands may simulta-neously offer the same trivial attribute, thereby reducing thebenefit of uniqueness (Brown and Carpenter 2000). Namely,consumers who make choices between high or low equitytier alternatives, in which one alternative offers a trivialattribute, may use information about whether a midequityalternative also offers the trivial attribute as an additionalinput to infer its value. We expect that this context cue willinteract with brand equity to influence consumer valuationof the trivial attribute.


Unique context. The choice context in which the trivialattribute is uniquely offered reinforces the brand cue’s valu-ation of the trivial attribute. Specifically, consumers expecthigh equity brands to offer unique advantages over mid- andlow equity brands. However, consumers do not expect a lowequity brand to offer a unique advantage over the higherequity of a midtier brand. Thus, we predict that a high equitybrand will be more likely than a low equity brand to gain anadvantage over its intratier competitor from a unique trivialattribute in the face of a prechoice disclosure. This predic-tion is counter to Carpenter, Glazer, and Nakamoto’s (1994)price result, which shows that a prechoice disclosure resultsin significant attenuation for a trivial attribute at the pre-mium price level.

Shared context. The preceding argument suggests a lim-ited ability for a low tier brand to initiate a trivial attributestrategy in the event of prechoice disclosure. However, a lowtier brand may be able to benefit within its competitive setfrom sharing a trivial attribute with a higher tier brand. Shar-ing a trivial attribute with a midequity brand is predicted toproduce opposite results from unique trivial differentiationfor low versus high equity brands. We expect that highequity brands offering a trivial attribute will suffer fromsharing the attribute with a lower equity brand. Sharing acommon attribute with a midequity brand simultaneouslyreduces its perceived differentiation from a lower equity tierand increases its high equity competitors’ perceived differ-entiation from the midequity tier. Conversely, we expect thata low equity brand sharing a trivial attribute with a higherequity brand will be viewed favorably. For a low equitybrand, the midtier brand represents a higher equity standard,and consumers are expected to believe that it offers superiorfeatures. Thus, in the face of a disclosure, we expect con-sumers’ choices to be positively influenced in the low tierwhen a target brand shares a trivial attribute with a midtierbrand. This prediction also runs counter to Carpenter,Glazer, and Nakamoto’s (1994) price result, which showsthat a low-priced alternative did not benefit from a trivialattribute strategy.

Therefore, we predict that consumer valuation of the triv-ial attribute is dependent on the interaction between thebrand equity tier and the decision context.

H1: A brand’s ability to gain a choice share benefit over a closecompetitor when the triviality of the attribute is disclosedprior to choice depends on its equity and the context, suchthat (a) a low equity brand will be more likely to gain a choice

share benefit if the trivial attribute is shared with a mid-equity brand than if it is uniquely offered and

(b) a high equity brand will be more likely to gain a choiceshare benefit if the trivial attribute is uniquely offeredthan if it is shared with a midequity brand.

STUDY 1

The purpose of Study 1 was to examine the effects of atrivial attribute strategy and the moderating roles of brandtier, disclosure of irrelevance, and decision context onchoice between intratier competitors. The choice setincluded three brands. Two of the brands were competitorsfrom the same equity tier, and the third was a midtier brand.

In all conditions but the control, a positive level of a trivialattribute was present for one of the two intratier brands (i.e.,the target brand), and a negative level was present for theother. The experiment examined two levels of brand tier(low and high), two levels of disclosure (subjective and pre-revelation), and two levels of decision context (unique andshared) for a 2 × 2 × 2 between-subjects experimentaldesign. The brand tier factor varied on whether the two com-petitive brands were from a low or a high equity tier. Thedisclosure factor varied on whether subjects were providedwith a prechoice disclosure that revealed that the trivialattribute was meaningless. In the condition in which the pre-choice disclosure was not provided, subjects determined thevaluation of the trivial attribute subjectively. The decisioncontext factor varied on whether the positive trivial attributelevel was unique to one of the intratier brands (unique con-text) or was also possessed by the midtier brand (shared con-text). In addition, there was a control condition for eachbrand tier in which subjects chose between branded alterna-tives without a trivial attribute.

The stimuli were based on those of Carpenter, Glazer, andNakamoto (1994) to increase comparability. The productcategory was down jackets, which were described by the rel-evant attributes of fill rating, cover material, and stitchingand the trivial attribute of type of down fill. To maximize thecorrespondence between the disclosure and the trivial attrib-ute (Broniarczyk and Gershoff 1997), the types of down fillwere described as either “goose” or “duck” down. A pretestof 50 undergraduate students using a nine-point rating scale(1 = “like” to 9 = “dislike”) revealed that goose was morepreferred than duck (x� = 3.88 versus x� = 5.40, t(49) = 3.83,p < .001), so goose and duck down were used as the positiveand negative levels of the trivial attribute, respectively.

Brands for the intratier pairs were selected through apretest in which 28 undergraduate students reported brandname evaluations on a nine-point scale (1 = “poor” to 9 =“excellent”). Store brands were used for this manipulationbecause our southwestern U.S. subjects were unfamiliarwith brands in the down jacket category. On the basis of theratings, we designated Kmart (x� = 3.11) and Wal-Mart (x� =3.89) as low tier brands, Mervyns (x� = 5.00) as the midtiercontext brand, and L.L. Bean (x� = 7.11) and Eddie Bauer(x� = 7.96) as high tier brands. All brands were highly famil-iar to subjects and received ratings greater than 8 on a nine-point scale. To avoid ceiling effects, the target brand pairedthe positive trivial attribute of goose down fill with the lowerrated brand in each tier (high: L.L. Bean and low: Kmart).The stimuli are presented in Appendix A.

Procedure

Three hundred forty-four undergraduate subjects enrolledin an introductory marketing class were given course extracredit for participating in Study 1. The cover story was thatthey needed to purchase a down jacket for an upcoming skivacation. Subjects received booklets containing a ConsumerReports description of down jackets on four attributes and,if appropriate, a disclosure manipulation embedded in thedescription of down fill type. Consistent with Carpenter,Glazer, and Nakamoto’s (1994) study, the disclosure manip-ulation informed subjects that the down fill attribute wasirrelevant with the phrase, “the age of the bird determines


4In the comprehension test, subjects answered questions about each ofthe four attributes of fill rating, stitching, down fill type, and cover material.For each attribute, the question began with the phrase, “According to theConsumer Reports information....” This was followed by four possibleresponses about the superiority of each attribute level: one or the otherattribute levels was superior, neither was superior, or there was not enoughinformation provided to answer the question. Subjects needed to answer allfour questions correctly to pass the comprehension test.

5Ninety percent of subjects in the low tier condition chose the midtiercontext brand, whereas none of the subjects in the high tier chose themidtier brand. None of the subjects’ rankings in either the high or the lowtier were inconsistent with their choice from the set of three brands.

the softness of the down fill, the type of bird it comes fromdoes not make a difference.” Subjects in the subjective con-dition were told only that two types of down fill were duckdown and goose down. To ensure that the prechoice disclo-sure manipulation was successful, subjects completed acomprehension test to indicate their knowledge of the attrib-ute descriptions before proceeding to the choice task.4

The choice task involved selecting the most preferredbrand from the set of three and then providing a ranking ofthe three brands. The measure of intratier choice was takenfrom subjects’ ranking of the two intratier alternatives inboth the low and high brand tier choice.5 Subjects then ratedbrand preference, familiarity, trust, and manufacturingexpertise. The last section assessed subjects’ beliefs aboutmarketer intent for offering the trivial attribute (Campbell1995). Specifically, the likelihood that the trivial attribute

was used to attract attention, improve product performance,or take advantage of the consumer was assessed on a five-point scale (1 = “very unlikely” to 5 = “very likely”) for theunique context.

Results

In examining the effects of trivial attribute strategiesbetween brand tiers, it is important to consider not only thedifferences in the brands’ choice shares between conditionsbut also the gain or loss in choice share relative to the con-trol condition in which there is no trivial attribute. There-fore, we performed a logit analysis, with preference for thetarget versus nontarget intratier brand as the dependent vari-able. We included independent variables in the model suchthat the effects of the manipulation of brand tier, disclosure,and context could be examined in terms of the change in thetarget brand choice share compared with its control. In Table1, we report absolute choice shares for each condition, andthe statistical analyses examine the difference (∆) betweenthe experimental and control conditions. Unless otherwisenoted, all χ2 have 1 degree of freedom.

Manipulation checks. The data were screened to ensurethat subjects correctly understood the Consumer Reportsattribute information and the disclosure regarding the trivialattribute. Analyses included only the 76% of subjects (n =260/344) who passed the prechoice disclosure comprehen-sion test. Consistent with the pretest, subjects reportedhigher means for the high than low tier brands on preference

Table 1STUDY 1 AND 1A CHOICE RESULTS

Low Equity Tier

Nontarget Brand: Target Brand:Wal-Mart Kmart Midtier Brand* Choice of Target Brand Down Fill Down Fill Down Fill (Standard Error)

Control — — — 4/34 12% (.06)

SubjectiveStudy 1 unique context Duck Goose Duck 11/22 50% (.11)Study 1 shared context Duck Goose Goose 15/24 63% (.10)Study 1A shared context Regular Alpine Alpine 20/26 77% (.08)

PrerevealedStudy 1 unique context Duck Goose Duck 7/26 27% (.09)Study 1 shared context Duck Goose Goose 12/20 60% (.11)Study 1A shared context Regular Alpine Alpine 15/24 63% (.09)

High Equity Tier

Nontarget Brand: Target Brand:Eddie Bauer L.L. Bean Midtier Brand* Choice of Target Brand Down Fill Down Fill Down Fill (Standard Error)

Control — — — 11/36 31% (.08)

SubjectiveStudy 1 unique context Duck Goose Duck 11/21 52% (.11)Study 1 shared context Duck Goose Goose 17/24 71% (.09)Study 1A shared context Regular Alpine Alpine 30/38 79% (.07)

PrerevealedStudy 1 unique context Duck Goose Duck 15/27 56% (.10)Study 1 shared context Duck Goose Goose 9/23 39% (.10)Study 1A shared context Regular Alpine Alpine 12/25 48% (.10)

*The midtier brand in Study 1 was Mervyns. Study 1A provided a set of four midtier brands: Sears, J.C. Penney, Mervyns, and Bealls.


(Mhigh = 7.25 versus Mlow = 4.84, t(258) = 13.331, p < .001),trust (Mhigh = 7.20 versus Mlow = 5.30, t(258) = 10.55, p <.001), and manufacturing expertise (Mhigh = 7.19 versusMlow = 4.76, t(258) = 14.18, p < .001).

Overall model. The choice results revealed no main effectfor share gain compared with the control by brand tier(∆high = +24.1% versus ∆low = +37.2%, χ2 = 1.90, p > .16)(see Table 1). There was a marginal effect of decision con-text; the target brand achieved slightly more share gain inthe shared (∆ = +36.8%) than in the unique (∆ = +24.4%)context (χ2 = 2.80, p = .09). As we expected, there was a sig-nificant main effect for disclosure; there was more sharegain for the target brand in the subjective (∆ = +37.9%) thanin the prerevelation (∆ = +23.4%) condition (χ2 = 3.57, p =.058). However, consistent with Carpenter, Glazer, andNakamoto (1994), even when prerevealed, the presence of atrivial attribute still resulted in a significant gain in choiceshare relative to the control (∆ = +44.8% versus +21.4%,χ2 = 9.87, p < .01).

Choice results. The results show a significant three-waybrand tier × context × disclosure interaction (χ2 = 3.71, p =.054). As we expected, in the subjective condition, there wasno significant brand tier × context interaction (χ2 < 1). In theabsence of a disclosure, there was a significant main effectfor adding a trivial attribute compared with the control (∆ =+37.9%, χ2 = 21.31, p < .001) but no effect for context(∆shared = +45.3% versus ∆unique = +28.6, χ2 = 2.25, p = .13)or brand tier (∆high = +31.6% versus ∆low = +47.4%, χ2 =1.52, p = .22). Without a disclosure, both high (χ2 = 7.73,p < .005) and low (χ2 = 15.04, p < .001) equity brandsgained choice share relative to the control. Subjectsappeared to rely on their prior preference for goose overduck down fill in valuing the trivial attribute and did not usethe context or brand tier to infer relative value to theattribute.

H1 predicts that when a prechoice disclosure reveals thatthe trivial attribute is meaningless, consumer valuation ofthe trivial attribute is dependent on the interaction betweenthe brand tier and the context. Specifically, we predicted thata low tier brand would gain choice share from its closestcompetitor through a trivial attribute strategy when it shared

the attribute with a midtier context brand. Conversely, wepredicted that a high tier brand would gain choice share in aunique context.

In support of H1, in the prerevelation condition, the pre-dicted brand tier × context interaction occurred (χ2 = 5.81,p < .02). As predicted by H1a, the low tier gained greaterchoice share when it shared (∆shared = +48%) the trivialattribute with a midtier context compared with when ituniquely offered the trivial attribute (∆unique = +15%, χ2 =4.88, p < .03). Furthermore, this share gain was significantrelative to the control in the shared (χ2 = 11.91, p < .001) butnot the unique (χ2 = 2.16, p = .14) context.

Consistent with H1b, when the irrelevance of the trivialattribute was disclosed, the high tier brand directionallygained more choice share from its competitor when ituniquely offered the trivial attribute than when the attributewas shared with a midtier context brand, though this differ-ence did not reach significance (∆unique = +25% versus∆shared = +8%, χ2 < 1). As we predicted, the high tier brand’sshare gain was significant relative to the control in theunique (χ2 = 3.89, p < .05) but not the shared (χ2 = .46, p =.49) condition (see Figure 1). Thus, when the disclosure wasprerevealed, subjects appeared to pay careful attention to thecontext brand and used its relative value to resolve theuncertainty that the disclosure created regarding the trivialattribute. After receiving a prechoice disclosure, subjectspreferred the high tier brand to be different from the midtierbrand but preferred the low tier brand to be similar to themidtier brand.

Consumer attributions. We assessed subjects’ beliefsabout the target brand’s motives for offering the trivialattribute (relative to scale midpoint) in the unique context toprovide insight into the consumer thought process and neg-ative reciprocal feedback to the brand. Subjects in the sub-jective conditions were neutral about whether the uniqueattribute improved performance (Mlow = 3.0 and Mhigh = 3.1,ps > .20) and believed that it was unlikely that the brand wastrying to take advantage of them (Mlow = 2.3 versus Mhigh =2.3, ps < .05). They believed that the reason the uniqueattribute was offered was to attract their attention (Mlow =3.6 and Mhigh = 3.7, ps < .05).

Figure 1STUDY 1 CHOICE SHARE FOR TARGET BRAND DIFFERENCES FROM CONTROL

0

10

20

30

40

50

60

0

10

20

30

40

50

60

Dif

fere

nce

fro

m C

on

tro

l

Prerevelation Subjective

Uniquecontext

Sharedcontext

High Equity Low EquityHigh Equity Low Equity


As we expected, when the trivial attribute was revealed tobe meaningless, consumers were significantly less likely tobelieve that the trivial attribute improved product perform-ance (with decline from the scale midpoint) in the low(Mlow = 2.4, p < .01) and high (Mhigh = 2.6, p < .10) tierswith no difference between tiers (t < 1, p > .20). Subjects inthe high tier were more likely to believe that the trivialattribute strategy was used as an attention-getting device(Mhigh = 4.4) compared with those in the low tier (Mlow =3.3, t(51) = 3.45, p < .001). There was little indication ofmanipulative intent by either high or low tier brands, in thatneither was greater than the midpoint of the scale (Mhigh =3.1, Mlow = 2.3, ps > .2), though the high tier was found tobe greater than the low tier brand (t(51) = 2.58, p < .02).

Summary. For a trivial attribute of subjective value, bothlow and high equity brands benefited from offering a trivialattribute that appeared favorable and was offered by otherbrands in the choice set, irrespective of their brand equity.This finding refutes a uniqueness explanation for the effectsof trivial attributes.

The presence of a disclosure appeared to make brandequity and context information more influential in the con-sumer evaluation process of the trivial attribute. In the pre-choice disclosure condition, prior beliefs about the trivialattribute were challenged by the disclosure. In this case, sub-jects were more likely to use other cues, such as brandequity and context brand offerings, to resolve the ambiguity.A high tier brand was more likely than a low tier brand tocontinue benefiting from a unique trivial attribute in the faceof a prechoice disclosure; subjects attributed it to anattention-getting device but did not ascribe manipulativeintent. However, when the trivial attribute was shared with amidtier brand, a low tier brand was more likely than a hightier brand to continue benefiting.

STUDY 1A

Prior research has focused on the benefits of uniquelyoffering a trivial attribute. Of particular interest in Study 1was the finding that when a trivial attribute is revealed to bemeaningless, a low but not high tier brand benefits from ashared context. Study 1A was designed to examine therobustness of a shared context for trivial attribute valuation.

In addition, we wanted to examine the effect of consumerknowledge on the valuation of the trivial attribute. Trivialattributes have included both attributes for which consumershave had prior preferences and knowledge and those forwhich they have not (Brown and Carpenter 2000). We mightexpect that consumers with low knowledge would be mostprone to value a trivial attribute positively. Yet even thebeliefs of high knowledge consumers regarding the trivialattribute might not be veridical (Alba and Hutchinson 2000).Therefore, we measure prior experience with the trivialattribute in Study 1A to examine its effect on valuation ondisclosure.

Finally, we instituted stimuli and procedural changes toaddress some deficiencies in the prior study. First, the trivialattribute continued to be down fill type, but the levels werenow described as “alpine” and “regular” rather than “goose”and “duck.” This served to duplicate the attribute labels usedby Carpenter, Glazer, and Nakamoto (1994), allowing for amore direct comparison with previous research. It also

served to examine a situation in which embellished labelsare used in place of actual product attribute descriptions,which may provide more latitude for inference making. Sec-ond, Study 1A moved from a simultaneous choice set to alearning and test phase that equated the procedure and stim-uli for the low– and high–brand equity condition. In Study1, subjects chose their most preferred option from a set ofthree alternatives (two intratier brands and a midtier brand)and then rank-ordered their preferences for the three alter-natives. Consequently, in the low tier, most subjects chosethe midtier brand, and preference between the two low tierbrands was measured by the rankings of three alternatives.Another potential problem was that the manipulation ofbrand tier confounded brand name and the performancelevel of the relevant attributes for the intratier brands.

Study 1A’s learning phase established the common levelof alpine fill material for midtier brands in the product cate-gory. Subjects read a Consumer Reports discussion of downjacket attribute information in which alpine down wasdescribed as made from goose feathers and regular downwas described as made from duck feathers. A pretest of 50undergraduate students using a nine-point rating scale (1 =“like” to 9 = “dislike”) confirmed that alpine was more pre-ferred than regular down fill (x� = 3.68 versus x� = 4.92,t(49) = 3.42, p < .001). In the prerevelation condition, Con-sumer Reports also disclosed the irrelevance of the type ofbird feather in determining down fill softness. Subjectscompleted a rating task of four midtier brands that all pos-sessed the shared positive level of the trivial attribute, alpinedown fill.

Subjects were then asked to choose between two intratiercompetitors that were comparable except for brand nameand the trivial attribute. The target brand is the brand offer-ing the positive shared trivial attribute of alpine down fill. Toprevent ceiling effects, the shared, more preferred alpinedown fill level was paired with the lower rated of the com-petitor brands (high: L.L. Bean and low: Kmart), and theunique, less preferred regular down fill level was paired withthe higher rated brand (high: Eddie Bauer and low: Wal-Mart). The attribute descriptions for the test brands wereidentical for the low and high tier conditions (see AppendixA).

Subjects were asked to describe the basis for their choice,followed by ratings of brand preference, familiarity, trust,and manufacturing expertise, as well as previous ownershipof a product containing down fill. All subjects were thenasked an open-ended question about the intent of the non-target brand offering the unique negative trivial attribute ofregular down fill.

One hundred thirteen subjects participated in exchangefor extra credit in an introductory marketing class. Theexperiment examined two levels of revelation (subjectiveand prerevealed) and two levels of brand tier (high and low)for a 2 × 2 between-subjects design. Only the shared contextwas examined. The control condition of Study 1 was used asthe baseline estimate for choice preference within a brandtier.

Results

Choice results. The results appear in Table 1. H1 predictsthat the ability of a brand using a trivial attribute to benefit


6Because we did not collect ownership measures in the control conditionin Study 1, only the data from Study 1A could be used. Comparisons arebased on actual differences of choice share without controlling for the dif-ference in share between the tiers.

over a close competitor when the triviality of the attribute isdisclosed depends on brand equity and context. Specifically,for the shared context, a low but not high tier brand isexpected to benefit from a trivial attribute strategy. Consis-tent with this hypothesis and corroborating the results ofStudy 1, there was a revelation × brand tier interaction rela-tive to the control (χ2 = 4.23, p < .05). Consistent with H1a,in the low tier when the target brand possessed alpine downfill, it gained significant share compared with the control inboth the subjective (∆ = +65%, χ2 = 20.72, p < .001) and therevealed (∆ = +51%, χ2 = 13.83, p < .001) conditions withno difference as a function of revelation (χ2 = 1.22, p = .27).That is, a low tier brand benefits from sharing the trivialattribute with a midtier brand, even when it is revealed thatthe attribute is meaningless. Consistent with H1b, in the hightier, there was a significant difference as a function of reve-lation (χ2 = 6.16, p < .02). The target brand possessingalpine down fill gained share when the trivial attribute wassubjective (∆ = +48%, χ2 = 15.87, p < .01), but there was nosignificant share gain when the attribute was revealed to bemeaningless (∆ = +17%, χ2 = 1.89, p = .17). Therefore, ahigh tier brand may benefit from sharing a trivial attributewith a midtier alternative only when consumers are unawareof the trivial attribute’s true value. However, when the con-sumer is aware of the true value of the trivial attribute, a hightier brand does not benefit.

Consumer attributions. To shed light on consumer reac-tion to a brand that does not share a trivial attribute, weassessed subjects’ cognitive responses in the prerevelationcondition regarding the intent of the sole brand to offer thenegative level of regular down fill. Subjects who received adisclosure were more likely to report a negative reaction tothat brand for failing to offer the shared level of the positivetrivial attribute of alpine for low than for high tier brands(Z = 3.57, p < .001). Specifically, 75% (18/24) of low brandtier subjects made negative inferences that the brand offer-ing regular instead of alpine down fill did so to take advan-tage of cost savings or ease of availability compared withonly 24% (6/25) of the high brand tier subjects. Thus, con-sistent with the choice results, subjects who were prein-formed of the irrelevance of the trivial attribute believed thata low tier brand should not be different from other categorymembers in offering the trivial attribute.

Next, we examined the generalizability of this resultacross different levels of prior experience with the trivialattribute. In Study 1A, 58% of subjects indicated that theyowned a product containing down fill, and these subjectswere evenly distributed across all cells such that no cell dif-fered significantly from the mean (ps >.1). A separate modelwas run to examine whether ownership affected choice ofthe target brand.6 In support of H1, the brand tier × revela-tion interaction was marginally significant (χ2 = 2.84, p =.092). The results also show a significant ownership × reve-lation interaction (χ2 = 3.71, p = .054), such that the effectof revelation was greater for nonowners than for owners ofdown fill products. However, no other significant maineffects or interactions were detected (ps > .15). Thus, the

7Specifically, Keller and Aaker (1992) find that the feedback from thesuccess of an intervening extension leads to more favorable consumer reac-tions to a proposed extension for an average quality core brand, whereas anunsuccessful intervening extension leads to less favorable consumer reac-tions for a high quality brand. They used hypothetical brands and providedsubjects with explicit information about the marketplace success of theextension. Although this research highlights the effect of prior marketingactions on future marketing actions, its brand results may have limited gen-eralizability to the situation examined here, in which real brands areexpected to influence the subjective valuation of the ambiguous trivialattribute.

finding that high tier brands can uniquely offer trivial differ-entiation, whereas low tier brands benefit only when thetrivial attribute is shared with a higher equity brand, appearsto generalize across prior experience with the trivialattribute.

Summary. We note that in both Study 1 and Study 1A,when subjects were unaware of the meaninglessness of thetrivial attribute, both low and high tier brands were able togain share from close competitors. Apparently, subjectsrelied on their prior beliefs to infer value from the attractivelevel of the trivial attribute. Thus, it would appear thatbrands have very little to lose and much to gain by using atrivial attribute strategy if consumers are unlikely to beaware of the true value of the trivial attribute at the time ofchoice. However, consumers may become aware of themeaninglessness of the trivial attribute after they have madea purchase decision. When this happens, consumers mayalter their view of the brands and perceive them as havingbeen manipulative by taking advantage of their equity with-out offering any additional value.

A critical test of brand dilution is how a brand’s priormarketing actions affect the success of its future marketingactions (Keller 1993). Specifically, we are interested inwhether postchoice revelation of the meaninglessness of thetrivial attribute will hurt the brand’s subsequent ability tointroduce a new differentiated attribute.

BRAND DILUTION: POSTCHOICE DISCLOSURE OFTRIVIAL DIFFERENTIATION

Prior research in which consumers received negativeinformation about a brand extension has generally shownonly limited dilution to overall brand attitude (Keller andAaker 1992) and brand beliefs (Loken and John 1993;Romeo 1991). This brand dilution research has beengrounded in categorization theory, whereby the feedback tothe core brand is a function of the similarity between theextension and original category and the degree to which theextension attributes are consistent with brand associations.The likelihood of dilution has been found to increase fornegative extensions to high similarity categories, such asline extensions (John, Loken, and Joiner 1998; Milberg,Park, and McCarthy 1997), that are the domain of trivialdifferentiation.

Brand dilution can be measured not only by examiningchanges to core brand beliefs and attitudes but also bywhether there is a change in consumer reaction to subse-quent brand actions. Keller and Aaker (1992) find that anunsuccessful intervening brand extension can negativelyaffect consumer reaction to a future brand extension.7 Thesuccess of intervening extensions affected the brand’s cred-ibility in introducing future extensions.


In our research, both high and low equity brands might beregarded as falsely implying that the trivial attribute is cred-ible through its mere existence (i.e., pragmatic inferences).Study 1 shows that subjects who received a prechoice dis-closure perceived low performance benefits from the trivialattribute. Yet these preinformed subjects made limited attri-butions of manipulative intent to the brand offering the triv-ial attribute. Instead, they appeared to rely on context brandinformation to resolve their uncertainty, preferring the triv-ially differentiated brand in the high tier when it wasuniquely offered and in the low tier when the trivial attributewas shared with a higher equity brand.

However, the extent of negative feedback may depend onthe timing of the disclosure. Prechoice disclosure providesinformation to the consumer that enables evaluation ofexpected outcomes. In many cases though, consumers maylearn of the irrelevance of the trivial attribute from sourcessuch as Consumer Reports, word of mouth, or competitoradvertising only after they have purchased the product. Con-sumers who were persuaded to choose on the basis of thetrivial attribute, only to discover later that it was meaning-less, may perceive the firm as having benefited without pro-viding similar benefit to the consumer. According to equityexchange theory, such a consumer comparison of the out-comes/inputs ratio relative to the brand’s outcomes/inputsratio would lead to a perception of unfairness (Campbell1995; Oliver and Swan 1989).

Equity exchange theory assumes an unbiased assessmentof the outcomes and inputs. But substantial research findsthat as a defense mechanism, consumers process informa-tion about a chosen alternative in a biased fashion (Eaglyand Chaiken 1993) by favorably interpreting ambiguousinformation (Muthukrishnan 1995) and generating counter-arguments to negative information (Ahluwalia, Burnkrant,and Unnava 2000). Such biased processing has been foundto be a function of brand equity. Brands for which con-sumers have higher commitment and stronger expectationsare more resistant to negative publicity (Ahluwalia,Burnkrant, and Unnava 2000) and product harm crises(Dawar and Pillutla 2000).

Because higher equity brands also possess higher levelsof brand trust (Chaudhuri and Holbrook 2001) and greatercertainty about new attribute performance (Nowlis andSimonson 1996), we may similarly expect choosers of highcompared with low tier brands to be less affected bypostchoice disclosure regarding the trivial attribute. As doKeller and Aaker (1992), we measure brand dilution byexamining how introducing a trivial attribute affects abrand’s subsequent ability to introduce a new differentiatedattribute.

On the basis of the preceding reasoning, we predict thatpostchoice disclosure will have a greater negative effect onsubsequent differentiation than will prechoice disclosureand that the postchoice disclosure will be more likely toaffect the consumers who originally chose the trivially dif-ferentiated brand. Furthermore, we predict that the effect ofthe postchoice disclosure on future differentiation isdependent on brand equity, in that high tier brands willreceive less negative feedback from a trivial attribute strat-egy than will low tier brands. We also examine consumerattributions as to marketer motives for offering the newattribute to gain insight into the potential negative feedback

to brand equity (Campbell 1999). Because negative attribu-tions are more likely for a brand that uniquely introducesrather than shares a trivial attribute with context alternatives,we focus on the context in which the brand uniquely offersthe trivial attribute.

H2: After postchoice disclosure of the irrelevance of a trivialattribute, original choosers of the brand differentiated by aunique trivial attribute will be less likely to choose thebrand’s subsequent new differentiated alternative in arelated product category for low than for high equity brands.

STUDY 2

Study 2 was a 2 (brand tier) × 2 (disclosure) between-subjects design with two sequential choices. The first factorof brand tier varied on whether the test brands were from alow or a high equity tier. The second factor of disclosure var-ied on whether the disclosure was provided to subjectsbefore the first choice (prerevelation) or after the first choicebut before the second choice (postrevelation). Only theunique context was examined; the target brand possessed theunique positive attribute of goose, and other brands pos-sessed duck down fill. In addition, there was a control con-dition for each brand tier in which subjects chose betweenbranded alternatives without a trivial attribute.

Procedure

Two hundred thirty-three subjects were given course extracredit for participating in Study 2. The cover story was thatthey needed to purchase a down jacket for an upcoming skivacation. The procedure was similar to Study 1A. In thelearning phase, subjects read a Consumer Reports discus-sion of down jacket attribute information. In the prerevela-tion condition only, Consumer Reports also disclosed theirrelevance of the type of bird feather in determining downfill softness, and subjects completed a comprehension test.The postrevelation condition was similar to the subjectivecondition in Study 1 through the first choice. Subjects thencompleted a rating task of four midtier brands that all pos-sessed the shared negative level of the trivial attribute, duckdown fill.

Subjects were then asked to choose between two intratiercompetitors that were comparable except for brand nameand the trivial attribute. The target brand paired the uniquepositive level of goose down fill with the lower rated of thecompetitor brands (high: L.L. Bean and low: Kmart). Theshared level of duck down fill was paired with the more pre-ferred brand (high: Eddie Bauer and low: Wal-Mart). Thestimuli appear in Appendix B.

Subjects were asked an open-ended question regardingthe basis for their choice, followed by ratings of brand pref-erence, familiarity, trust, and manufacturing expertise. Then,in the postrevelation condition only, subjects received anupdate from Consumer Reports with the trivial attributedisclosure.

The next section contained the second choice scenario.Subjects were told that their vacation plans had been post-poned until later in the spring, so instead of a heavy downjacket, they would need only a lighter fleece jacket for theski trip. Information was provided about three fleece jacketattributes of fleece fiber, weight, and water repellant. Thesecond choice task asked subjects to select between the


same two intratier brands as were provided in the firstchoice. The alternatives were comparable on the attributesof fleece fiber, weight, pockets, zipper, and color options.The differentiation existed on ambiguous types of the waterrepellant attribute. The target brand (low: Kmart and high:L.L. Bean) that had offered the unique trivial attribute in thedown jacket choice now offered a “NEW!” PKX waterrepellant in the fleece jacket category, whereas the nontargetbrand offered BXZ water repellant. The highlighting of“NEW!” for the target brand was designed to test subjectreaction to the brand’s latest differentiation that was neutralin value.

After making a choice, subjects were asked to describethe basis for their choice, preference for water repellanttype, and beliefs about marketer intent for offering the newtype of water repellant. Specifically, we assessed the likeli-hood that the “NEW!” PKX water repellant was used toimprove product performance, attract attention, or takeadvantage of the consumer on a five-point scale (1 = “veryunlikely” to 5 = “very likely”).

Results

The data were screened to ensure that subjects correctlyunderstood the Consumer Reports attribute information andthe disclosure regarding the trivial attribute. Analysesincluded only the 81% of subjects (n = 188/233) who passedthe prechoice comprehension test. As confirmation of thepretest, subjects reported higher means for the high than forthe low tier brands on preference (x�high = 6.28 versus x�low =5.47, t(186) = 2.15, p < .033), trust (x�high = 6.53 versusx�low = 5.96, t(186) = 2.86, p < .005), and manufacturingexpertise (x�high = 6.25 versus x�low = 5.08, t(185) = 5.74, p <.001).

Choice 1 results: down jackets. We used a logit modelwith choice of the target brand as the dependent variable andbrand tier and disclosure as independent variables to test fordifferences in choice share compared with the control. Thepostrevelation disclosure condition was identical to the sub-jective condition in Study 1 for the subjects’ first choice(down jackets).

The choice results revealed no main effect for target brandchoice share by brand tier (x�high = 37.2% versus x�low =29.0%, χ2 = 1.41, p = .23). As we expected, there was a sig-nificant main effect for disclosure with more share gain forthe target brand in the postrevelation (∆ = +32.4%) than the

prerevelation (∆ = +10.1%) condition (χ2 = 5.43, p < .02).When a disclosure prerevealed the triviality of the attribute,there was a gain in choice share relative to the control, but itdid not attain significance (x�prerevelation = 30.4% versusx�control = 20.3%, χ2 = 1.72, p = .19).

H1 suggests that a high tier brand will be more likely thana low tier brand to maintain its choice share benefit on dis-closure when the trivial attribute is uniquely offered.Although the disclosure × brand tier (χ2 = .436, p < .5) inter-action was not significant, the analysis shows significantattenuation from a prechoice disclosure in the low tier(∆postrevelation = +31.1% versus ∆prerevelation = +4.4%, χ2 =4.16, p < .05) but not in the high tier (∆postrevelation = +33.2%versus ∆prerevelation = +16.9%, χ2 = 1.51, p > .22). However,the gain in the high tier on revelation did not attain signifi-cance (p = .17) (see Table 2).

Choice 2 results: fleece jackets. The purpose of thepostrevelation condition was to examine how a targetbrand’s customers might react to learning of the irrelevanceof the trivial attribute after they had made their choice. H2predicts that postchoice disclosure will make originalchoosers of the trivial differentiation less likely to choosethe new differentiation in the low than in the high brand tier.We used a logistic regression model to examine subjects’second choices (fleece jacket choice) as a function of thedisclosure condition, the brand tier, and subjects’ priorchoices (down jacket choice).

Not surprisingly, there was a main effect from subjects’first brand choice (χ2 = 29.95, p <.001). Subjects who chosethe target brand with the trivial attribute in the first choicewere more likely to choose the target brand in the secondchoice (83.9%) than were those who had not chosen the tar-get in the first choice (32.3%). However, the tendency tochoose the target brand in the second choice also dependedon the brand tier and the disclosure conditions.

In support of H2, in the postrevelation condition com-pared with the control, the percentage of subjects whoselected the target brand in the second choice differeddepending on the brand tier and whether subjects had cho-sen the target brand in the previous choice (first choice ×brand tier: χ2 = 3.63, p =.057). In the control condition,100% of subjects who selected the target brand in the firstchoice selected it again when it offered a new attribute inboth the low and the high tier. In the postrevelation condi-tion, subjects in the low tier (53.8%) were less likely than

Table 2STUDY 2 CHOICE RESULTS

Second Choice (Fleece Jackets) Target Brand Choosers

First Choice (Down Jackets) First Choice First ChoiceTarget Brand Choosers Target Brand Choosers Nontarget Brand Choosers

Low TierControl 7/37 19% (.06) 7/7 100% 6/29 21% (.08)Postrevelation 13/26 50% (.10) 7/13 54% (.14) 5/13 39% (.14)Prerevelation 7/30 23% (.08) 5/7 71% (.17) 7/23 30% (.10)

High TierControl 8/37 22% (.07) 8/8 100% 12/29 41% (.09)Postrevelation 17/31 55% (.09) 16/17 94% (.06) 4/14 29% (.12)Prerevelation 10/26 39% (.10) 9/10 90% (.10) 6/16 38% (.12)

Notes: Standard errors are in parentheses.


Figure 2STUDY 2 SECOND CHOICE: TARGET BRAND’S SHARE OF

CHOICE 1 RETAINED IN SUBSEQUENT CHOICE OF NEW

BRAND ATTRIBUTE

0%

20%

40%60%

80%100%

Control Postrevelation

Highequity

Lowequity

those in the high tier (94.1%) to choose the target brand inthe second choice if they had chosen it in the first choice(χ2 = 7.043, p < .01) (see Figure 2). Thus, when disclosureregarding the trivial attribute occurred postchoice, a high tierbrand retained its share of choosers for the brand’s new dif-ferentiation, but the low tier lost a significant portion of itsoriginal choosers.

There was no significant difference by tier for postrevela-tion subjects who had not chosen the target brand in the firstchoice (x�low = 38.5%, x�high = 28.6%, χ2 = .297, p = .585). Inthe prerevelation condition when disclosure was revealedbefore the first choice, there was no significant tier by firstchoice interaction (χ2 = .711, p = .399).

When subjects were aware of the meaninglessness of atrivial attribute before a first choice, brand tier did notappear to affect the choice of an extension product by thosewho either had or had not chosen the target brand in theirfirst choice. However, subjects who chose a brand with atrivial attribute and subsequently learned that the differenti-ation was meaningless were less likely to choose an exten-sion of that brand in the low compared with the high brandtier.

Consumer attributions. Cognitive responses describingthe basis for the second choice revealed that the likelihoodof postrevelation subjects spontaneously reporting negativeinferences regarding the target brand’s new attribute (e.g.,

“New doesn’t always mean better,” “I am wary of the newwater repellent”) was significantly higher for low than forhigh tier brands. Among target brand choosers in the firstchoice, half of the low tier subjects (50.0%) made negativecomments about its offering a new attribute in the secondchoice product category, whereas few high tier subjects(5.9%) made negative comments about the new attribute(χ2 = 7.48, p < .01). However, there was no difference in thelikelihood of negative inferences for those who had not cho-sen the target brand in the first choice by brand tier (x�low =21.4% versus x�high = 28.6%, χ2 < 1). This first choice ×brand tier interaction for negative inferences was significant(χ2 = 4.59, p < .05). Thus, consistent with their secondchoices, subjects in the low tier were more likely to ascribea negative value to the new attribute in a subsequent choicewhen disclosure came after making a first choice, particu-larly those who had originally chosen the target brand.

Postrevelation subjects’ beliefs about the marketer’smotive for offering the new attribute in the second choiceprovide further insight (see Table 3). Consistent withcognitive responses, postrevelation caused subjects whochose the target brand in the first choice to be less likely (rel-ative to control subjects) to believe that the new attributeimproved performance in the low tier (x�control = 3.72 versusx�postrevelation = 2.69, F(1, 18) = 3.54, p < .08) but not in thehigh tier (x�control = 3.25 versus x�postrevelation = 3.24, F(1,23) = 0, p > .98) condition. All subjects strongly believedthat the brand offered the new attribute to attract attention(M = 4.65), which created a ceiling effect with no differ-ences from the control as a function of brand tier or firstchoice (ps > .20).

As we expected, attributions of taking advantage of theconsumer were directionally stronger for postrevelation sub-jects than for their counterparts in the prerevelation condi-tion (x�postrevelation = 4.14 versus x�prerevelation = 3.73, t(111) =1.79, p = .077). Although first choice target brand choosersin the postrevelation condition were directionally morelikely than control subjects to attribute manipulative intentto a brand attempting new differentiation in the low(x�control = 3.29 versus x�postrevelation = 3.85, F(1, 18) = .732,p > .40) and high (x�control = 4.00 versus x�postrevelation = 4.18,F(1, 23) = .16, p > .69) brand tiers, neither result was sig-nificant, perhaps because of small sample size. The highest

Table 3STUDY 2 SECOND CHOICE (FLEECE JACKETS) BELIEF MEASURES

Improve Performance Attract Attention Take Advantage

First Choice First Choice First Choice First Choice First Choice First ChoiceTarget Nontarget Target Nontarget Target Nontarget

Choosers Choosers Choosers Choosers Choosers Choosers

Low TierControl 3.71 (.42) 2.76 (.22) 4.71 (.18) 4.69 (.11) 3.29 (.61) 3.86 (.27)Postrevelation 2.69 (.33) 2.46 (.35) 4.77 (.12) 4.54 (.18) 3.85 (.36) 4.00 (.38)Prerevelation 3.14 (.46) 2.57 (.23) 4.43 (.30) 4.57 (.22) 3.29 (.64) 3.61 (.30)

High TierControl 3.25 (.67) 3.57 (.17) 4.75 (.25) 4.72 (.08) 4.00 (.33) 3.93 (.18)Postrevelation 3.24 (.33) 2.79 (.24) 4.65 (.12) 4.64 (.13) 4.18 (.26) 4.50 (.23)Prerevelation 3.40 (.50) 3.44 (.24) 5.00 (.00) 4.38 (.18) 4.20 (.36) 3.81 (.23)

Notes: Beliefs assessed the likelihood of the reason a brand offered a unique attribute on a five-point scale, where 1 = “very unlikely” and 5 = “very likely.”Standard errors are in parentheses.


ratings of manipulative intent for new differentiation werefor nontarget brand choosers in the high tier postrevelationcondition (x�control = 3.93 versus x�postrevelation = 4.50,F(1, 41) = 3.54, p = .067), in which the postchoice disclo-sure confirmed their decision to ignore the trivial attribute.

Summary. To summarize, the first choice in Study 2revealed a pattern similar to Study 1; prerevelation signifi-cantly attenuated the share gain from offering a unique triv-ial attribute for low but not high tier brands, though the hightier gain was limited. When disclosure occurred after sub-jects had made a choice between brands distinguished by atrivial attribute, the brands’ subsequent ability to introduce adifferentiated attribute was affected. Subjects in the low tiercondition who had selected the target brand in their firstchoice were far less likely to select the brand again (Choice2) than were those in the high tier. High tier brands dis-played almost no adverse effects on subsequent choice frompostrevelation of the trivial attribute’s value. After apostchoice disclosure, original choosers in the low but notthe high tier made negative inferences about the brand’s newattribute in a subsequent choice. Postchoice revelation madeall subjects believe that the target brand offered the newattribute as an attention-getting device and to take advantageof them, but this did not necessarily negatively influencetheir likelihood of choosing the target brand.

DISCUSSION

In summary, although brands commonly use trivial attrib-ute strategies as a means to lure consumers away from com-petitors, only recently have marketing researchers begun tounderstand the mechanisms and limitations of such strate-gies. Previous research has attributed the effectiveness oftrivial attribute strategies in the face of the disclosure ofmeaninglessness to consumer attraction to the uniqueness ofthe differentiated attribute (Carpenter, Glazer, andNakamoto 1994) and consumers’ use of the trivial attributefor decision resolution (Brown and Carpenter 2000). Wesuggest that brand equity, the choice context, and the timingof the disclosure of meaninglessness also influence theeffectiveness of trivial attribute strategies.

We found that brand equity and the offering of the trivialattribute by a context brand influenced the ability of a brandto gain choice share when the triviality of the differentiationwas disclosed before subjects’ choices. A low equity brandwas able to gain choice share when it differentiated itselffrom its closest competitor by sharing the trivial attributewith a higher equity context brand. Our results for highequity brands were less conclusive but generally showedthat a high equity brand was able to gain choice share whenit differentiated itself from both its closest competitor and alower equity context brand by uniquely offering a trivialattribute.

Without a disclosure, both high and low equity brandsincreased their choice share over their closest competitorthrough a trivial attribute strategy with limited contexteffects. When the meaninglessness of the trivial attributewas disclosed after subjects’ choices, choices of a subse-quent differentiation by the same brand were dependent onbrand equity. Low equity brands lost a substantial portion oftheir customer base, which made negative inferences aboutthe performance value of the subsequent new differentiation.Conversely, high equity brands retained their customer base,

which continued to select the brand’s new differentiatedoffering despite the postchoice revelation that the brand’sprior differentiation was meaningless.

This research adds to the literature on product entry bysuggesting that the success of an innovation strategy by alate mover (Shankar, Carpenter, and Krishnamurthi 1998) isdependent on brand equity. A strategic implication of ourresults is that brand equity affects a brand’s ability to main-tain its choice share in the face of trivial differentiation byother members of its product category. Low tier brands maybe penalized if they fail to offer the trivial differentiation ofhigher equity brands. In contrast, high tier brands that seekto offer only meaningful differentiation to maintain theircredibility might not be penalized if they refuse to followtrivial differentiation by other category members.

Our studies do not explicitly manipulate the temporalsequence of information presentation. Additional researchshould examine the dynamic valuation of a trivial attributeas its adoption diffuses over a product category. For exam-ple, the lower equity Walgreens’ store brand imitated thehigher equity Pantene Pro-V vitamin shampoo with its ownV vitamin shampoo. Our results suggest that as long as con-sumers are unaware of the meaninglessness of vitamins inshampoo, Pantene’s market share with respect to otherhigher equity brands will not suffer from this imitation. Wal-greens will also benefit by taking share from other lower tierbrands. However, as consumers become informed that vita-mins are a trivial attribute for shampoo (Consumer Reports2000), our results suggest that Pantene is less likely to con-tinue to benefit from the trivial attribute strategy but maystill be able to benefit from subsequent trivial differentia-tion. Walgreens may find that its customers are likely to bemore skeptical about future trivial differentiation. However,repeated experience with the trivial attribute may lead con-sumers to develop a preference for the meaningless attribute(Muthukrishnan and Kardes 2001) and possibly positivefeedback to the brand.

Study 1A does not show a moderating effect of priorownership of a down fill product on consumer valuation oftrivial attributes. However, our measure of product familiar-ity was a single-item question of prior ownership. Becauseproduct knowledge is a multifaceted construct, researchshould further examine the moderating effect of multipletypes of knowledge, such as category, trivial attribute, andbrand knowledge. Our results show that those with priorownership of a down product are more resistant to the dis-closure that the attribute of down fill type was trivial.Because consumers with high category knowledge are alsolikely to possess well-defined brand preferences, they maysimilarly exhibit high resistance to a disclosure if theirfavorite brand is explicitly promoting the trivial attribute(Ahluwalia, Burnkrant, and Unnava 2000).

Further research might also examine other cumulativeeffects of trivial attributes on the potential brand dilution ofhigh equity brands. Consumer experience with a trivialattribute may inadvertently block the learning of futurebrand attributes that offer meaningful differentiation (vanOsselaer and Alba 2000). In addition, revelation of the truevalue of the trivial attribute was shown to increase subjects’beliefs about the brand’s manipulative intent in both the lowand the high tier. Our results show that consumers ascribemanipulative intent to high tier brands after a postchoice dis-


closure that its differentiation was trivial but continue tochoose the brand’s future differentiations. It appears that aone-time negative encounter may not be enough to under-mine the sum total of positive brand experiences that a hightier may have accumulated in its equity. Although con-sumers’ persuasion knowledge may lead them to expectmarketers to engage in some manipulative tactics (Friestadand Wright 1994), continued use of a trivial differentiationstrategy may produce noticeable brand dilution, especially ifit is coupled with a price increase (Campbell 1999).Although our focus was on brand dilution among priorbrand choosers, the belief results of Study 2 also suggestthat dilution occurs for nonchoosers of a brand. Disclosureof the meaninglessness of the trivial differentiation madenonchoosers highly likely to ascribe manipulative intent tothe differentiated brand, which potentially solidifies theirpreference for a competitor.

It is also important to examine other components of brandequity. Our measure of brand equity was the overall evalua-tive rating assigned to the brand. Furthermore, all the brandsexamined were familiar to subjects. Recent research sug-gests that brand commitment (Ahluwalia, Burnkrant, andUnnava 2000), credibility (Erdem and Swait 1998), and trust(Chaudhuri and Holbrook 2001) are important brand equitycomponents to consider in addition to brand affect. Highequity brands may possess stronger consumer commitmentthan low equity brands, and this perhaps limited share gainfrom a trivial attribute strategy in the high tier following

prechoice disclosure. The signaling perspective of brandequity posits that the sunk costs invested in prior periodscreate a brand’s reputation that its claims are credible(Erdem and Swait 1998). Even our low equity brands hadestablished reputations. Additional research should examinethe possible interactions between brand affect and brandcredibility on the long-term success of a trivial differentia-tion strategy. Although manipulation checks showed that ourbrand affect and brand trust measures were correlated (r =.642, p < .001), it would be interesting to disentangle theseeffects. For familiar brands, there may be brand-specificassociations that can be leveraged as a basis for trivial dif-ferentiation (Kraus and Carpenter 1999). Prior research isequivocal regarding whether negative extension informationabout general equity or brand-specific association is morelikely to dilute the core brand (John, Loken, and Joiner1998).

This research extends understanding of the effects of triv-ial differentiation strategies by demonstrating the moderat-ing effect of brand equity and context. When triviality is dis-closed, low equity brands can benefit, but only if they sharethe differentiation with a higher equity brand, whereas highequity brands benefit, but only if they uniquely offer the dif-ferentiation. Finally, the effects of brand equity and trivialdifferentiation are reciprocal. Disclosure of trivial differen-tiation after a consumer makes a choice can affect a brand’sequity, such that subsequent trivial differentiation strategiesmay be less effective.

Appendix ASTIMULI FOR STUDY 1

Nontarget Target Context

High Quality TierEddie Bauer L.L. Bean Mervyns

550 fill rating 550 fill rating 500 fill ratingSynthetic cover Synthetic cover Synthetic coverExtra tight stitching Extra tight stitching Regular stitchingDuck down fill Goose down fill Goose/duck down fill

Low Quality TierWal-Mart Kmart Mervyns

450 fill rating 450 fill rating 500 fill ratingSynthetic cover Synthetic cover Synthetic coverSub-regular stitching Sub-regular stitching Regular stitchingDuck down fill Goose down fill Goose/duck down fill

STIMULI FOR STUDY 1A

Part I: Learning Phase—Down Jacket

Fill Cover Stitches RatingBrand Rating Down Fill Material per Inch (1–7)

Sears 500 Alpine Synthetic 12 ____J.C. Penney 525 Alpine Cotton 8 ____Mervyns 550 Alpine Cotton 12 ____Bealls 525 Alpine Synthetic 8 ____

Part II: Test Phase—Down Jacket (High Tier/Low Tier)

Eddie Bauer/Wal-Mart L.L. Bean/Kmart525 fill rating 525 fill ratingSynthetic cover Synthetic cover12 stitches per inch 12 stitches per inchDetachable hood Detachable hoodDry clean only Dry clean onlyRegular down fill Alpine down fill


Appendix BSTIMULI FOR STUDY 2

Part I: Learning Phase—Down Jacket

Fill Cover Stitches RatingBrand Rating Down Fill Material per Inch (1–7)

Sears 500 Duck Synthetic 12 ____J.C. Penney 525 Duck Cotton 8 ____Mervyns 550 Duck Cotton 12 ____Bealls 525 Duck Synthetic 8 ____

Part II: Choice 1—Down Jacket (High Tier/Low Tier)

Eddie Bauer/Wal-Mart L.L. Bean/Kmart525 fill rating 525 fill ratingSynthetic cover Synthetic cover12 stitches per inch 12 stitches per inchDetachable hood Detachable hoodDry clean only Dry clean onlyDuck down fill Goose down fill

Part III: Choice 2—Fleece Jacket (High Tier/Low Tier)

Eddie Bauer/Wal-Mart L.L. Bean/KmartPolartec fleece Polartec fleece200 weight 200 weightLined pockets Lined pocketsFull zipper Full zipperMultiple colors Multiple colorsBXZ water repellant NEW! PKX water repellant

REFERENCES

Ahluwalia, Rohini, Robert E. Burnkrant, and H. Rao Unnava(2000), “Consumer Response to Negative Publicity: The Moder-ating Role of Commitment,” Journal of Marketing Research, 37(May), 203–14.

Alba, Joseph W. and Wesley Hutchinson (2000), “Knowledge Cali-bration: What Consumers Know and What They Think TheyKnow,” Journal of Consumer Research, 27 (September), 123–56.

Allison, Ralph and Kenneth P. Uhl (1964), “Influences of BeerBrand Identification on Taste Perception,” Journal of MarketingResearch, 2 (August), 36–39.

Boulding, William and Amna Kirmani (1993), “A Consumer-SideExperimental Examination of Signaling Theory: Do ConsumersPerceive Warranties As Signals of Quality,” Journal of Con-sumer Research, 20 (June), 111–23.

Broniarczyk, Susan M. and Joseph W. Alba (1994), “The Role ofConsumers’ Intuitions in Inference Making,” Journal of Con-sumer Research, 21 (December), 393–407.

——— and Andrew D. Gershoff (1997), “Meaningless Differenti-ation Revisited,” in Advances in Consumer Research, Vol. 24,Merrie Brucks and Debbie MacInnis, eds. Provo, UT: Associa-tion for Consumer Research, 223–28.

Brown, Christina L. and Gregory S. Carpenter (2000), “Why Is theTrivial Important? A Reasons-Based Account for the Effects ofTrivial Attributes on Choice,” Journal of Consumer Research, 26(March), 372–85.

Buchanan, Lauranne, Carolyn J. Simmons, and Barbara A. Bickart(1999), “Brand Equity Dilution: Retailer Display and ContextBrand Effects,” Journal of Marketing Research, 36 (August),345–55.

Campbell, Margaret C. (1995), “When Attention-Getting Advertis-ing Tactics Elicit Consumer Inferences of Manipulative Intent:The Importance of Balancing Benefits and Investments,” Jour-nal of Consumer Psychology, 4 (3), 225–54.

——— (1999), “Perceptions of Price Unfairness: Antecedents andConsequences,” Journal of Marketing Research, 36 (May), 187–99.

Carpenter, Gregory S., Rashi Glazer, and Kent Nakamoto (1994),“Meaningful Brands from Meaningless Differentiation,” Journalof Marketing Research, 31 (August), 339–50.

Chaudhuri, Arjun and Morris B. Holbrook (2001), “The Chain ofEffects from Brand Trust and Brand Affect to Brand Perfor-mance: The Role of Brand Loyalty,” Journal of Marketing, 65(April), 81–93.

Consumer Reports (2000), “Head Games: Some Shampoo ClaimsAre Just a Lot of Lather,” 65 (September), 18–21.

——— (2001), “Scanners Get Cheap,” 66 (May), 24–27.Dawar, Niraj and Madan M. Pillutla (2000), “Impact of Product-

Harm Crises on Brand Equity: The Moderating Role of Con-sumer Expectations,” Journal of Marketing Research, 37 (May),215–26.

Ditto, Peter H. and David F. Lopez (1992), “Motivated Skepticism:Use of Differential Decision Criteria for Preferred and Nonpre-ferred Conclusions,” Journal of Personality and Social Psychol-ogy, 63 (April), 568–84.

Eagly, Alice A. and Shelly Chaiken (1993), The Psychology of Atti-tudes. New York: Harcourt Brace Jovanovich, Inc.

Erdem, Tulin and Joffre Swait (1998), “Brand Equity As a Signal-ing Phenomenon,” Journal of Consumer Psychology, 7 (2), 131–58.

Feldman, Jack M. and John G. Lynch (1988), “Self-GeneratedValidity and Other Effects of Measurement on Belief, Attitude,Intention, and Behavior,” Journal of Applied Psychology, 73 (3),421–35.

Friestad, Marian and Peter Wright (1994), “The Persuasion Knowl-edge Model: How People Cope With Persuasion Attempts,”Journal of Consumer Research, 21 (June), 1–31.

Gilbert, Daniel T. (1991), “How Mental Systems Believe,” Ameri-can Psychologist, 46 (February), 107–19.

Goldberg, Marvin E. and Jon Hartwick (1990), “The Effects ofAdvertiser Reputation and Extremity of Advertising Claims onAdvertising Effectiveness,” Journal of Consumer Research, 17(September), 172–79.


Houston, Michael J., Terry L. Childers, and Susan E. Heckler(1987), “Picture-Word Consistency and the Elaborative Process-ing of Advertisements,” Journal of Marketing Research, 24(November), 359–69.

Johar, Gita Venkataramani (1996), “Intended and UnintendedEffects of Corrective Advertising on Beliefs and Evaluations: AnExploratory Analysis,” Journal of Consumer Psychology, 5 (3),209–30.

——— and Carolyn J. Simmons (2000), “The Use of ConcurrentDisclosures to Correct Invalid Inferences,” Journal of ConsumerResearch, 26 (March), 307–22.

John, Deborah Roedder, Barbara Loken, and Christopher Joiner(1998), “The Negative Impact of Extensions: Can FlagshipProducts Be Diluted,” Journal of Marketing, 62 (January), 19–32.

Kahneman, Daniel, Jack L. Knetsch, and Richard Thaler (1986),“Fairness As a Constraint on Profit Seeking: Entitlements in theMarket,” The American Economic Review, 76 (4), 728–41.

Keller, Kevin Lane (1993), “Conceptualizing, Measuring, andManaging Customer-Based Brand Equity,” Journal of Market-ing, 57 (January), 1–22.

——— and David A. Aaker (1992), “The Effects of SequentialIntroduction of Brand Extensions,” Journal of MarketingResearch, 29 (February), 35–50.

Kraus, Paul J. and Gregory S. Carpenter (1999), “CompetitiveBrand Differentiation: What Makes Differences Valuable?”working paper, J.L. Kellogg Graduate School, NorthwesternUniversity.

Loken, Barbara and Deborah Roedder John (1993), “DilutingBrand Beliefs: When Do Brand Extensions Have a NegativeImpact?” Journal of Marketing, 57 (July), 71–84.

Milberg, Sandra J., C. Whan Park, and Michael S. McCarthy(1997), “Managing Negative Feedback Effects Associated withBrand Extensions: The Impact of Alternative Branding Strate-gies,” Journal of Consumer Psychology, 6 (2), 119–40.

Mukherjee, Ashesh and Wayne D. Hoyer (2001), “The Effect ofNovel Attributes on Product Evaluation,” Journal of ConsumerResearch, 28 (December), 462–72.

Muthukrishnan, A.V. (1995), “Decision Ambiguity and IncumbentBrand Advantage,” Journal of Consumer Research, 22 (June),98–109.

——— and Frank R. Kardes (2001), “Persistent Preferences forProduct Attributes: The Effects of Initial Choice Context andUninformative Experience,” Journal of Consumer Research, 28(June), 89–104.

Nowlis, Stephen M. and Itamar Simonson (1996), “The Effect ofNew Product Features on Brand Choice,” Journal of MarketingResearch, 33 (February), 36–47.

——— and ——— (1997), “Attribute-Task Compatibility As aDeterminant of Consumer Preference Reversals,” Journal ofMarketing Research, 34 (May), 205–18.

Oliver, Richard L. and John E. Swan (1989), “Equity and Discon-firmation Perceptions As Influences on Merchant and ProductSatisfaction,” Journal of Consumer Research, 16 (December),372–83.

Park, C. Whan, Sandra Milberg, and Robert Lawson (1991), “Eval-uation of Brand Extensions: The Role of Product Level Similar-ity and Brand Concept Consistency,” Journal of ConsumerResearch, 18 (September), 185–93.

Park, Chan Su and V. Seenu Srinivasan (1994), “A Survey-BasedMethod for Measuring and Understanding Brand Equity andExtendibility,” Journal of Marketing Research, 21 (May), 271–88.

Prelec, Drazen, Birger Wernerfelt, and Florian Zettelmeyer (1997),“The Role of Inference in Context Effects: Inferring What YouWant from What Is Available,” Journal of Consumer Research,24 (June), 118–25.

Romeo, Jean B. (1991), “The Effect of Negative Information onthe Evaluations of Brand Extensions and the Family Brand,” inAdvances in Consumer Research, Vol. 18, R.H. Holman and R.Solomon, eds. Provo, UT: Association for Consumer Research,399–406.

Shankar, Venkatesh, Gregory S. Carpenter, and Lakshman Krish-namurthi (1998), “Late Mover Advantage: How Innovative LateEntrants Outsell Pioneers,” Journal of Marketing Research, 35(February), 54–70.

Shapiro, Carl (1982), “Consumer Information, Product Quality,and Seller Reputation,” Bell Journal of Economics, 13 (1), 20–35.

Simonson, Itamar, Ziv Carmon, and Suzanne O’Curry (1994),“Experimental Evidence on the Negative Effect of Product Fea-tures and Sales Promotions on Brand Choice,” Marketing Sci-ence, 13 (Winter), 23–40.

———, Stephen Nowlis, and Katherine Lemon (1993), “TheEffect of Local Consideration Sets on Global Choice BetweenLower Price and Higher Quality,” Marketing Science, 12 (Fall),357–77.

———, ———, and Yael Simonson (1993), “The Effect of Irrele-vant Arguments on Consumer Choice,” Journal of ConsumerPsychology, 2 (3), 287–306.

van Osselaer, Stijn M.J. and Joseph W. Alba (2000), “ConsumerLearning and Brand Equity,” Journal of Consumer Research, 27(June), 1–16.

Date post:	26-Jan-2017
Category:	Documents
Upload:	nguyenxuyen
View:	225 times
Download:	2 times

The Reciprocal Effects of Brand Equity and Trivial Attributes

Documents